John R. Taylor

Cognitive Approaches to Lexical Semantics

This book provides a representative survey of early and more recent concerns in cognitively inspired lexical semantics. As such, it focuses on the issue of polysemy vs. monosemy, it offers fresh perspectives on prototypicality in lexical categories, it sheds light on the development of lexical items in child language acquisition and in diachrony, and it looks at issues going beyond the individual lexical item (onomasiology, synonymy, the relationship between lexical and syntactic meaning).

Near synonyms as co-extensive categories: ‘high’ and ‘tall’ revisited

Abstract One of the least understood semantic relations is synonymy. While perfect synonyms are rare, “near synonyms” are especially numerous. Near synonyms are words which are similar in meaning, which tend not to be contrastive, but which are distributed differently. Following up an earlier study of the adjective tall, this paper examines the near synonyms high and tall, and argues that the words offer different construals of an entitys verticality. Drawing on work by MacLaury in the domain of colour—in particular, MacLaurys discovery of what he calls “co-extensive” colour categories and his explanation of coextension in terms of a dominant and a recessive vantage—it is proposed that the distribution of the English adjectives can be insightfully analyzed in terms of the co-extension relation, with high designating the dominant vantage, tall the recessive vantage.

Polysemy's paradoxes

Abstract This article reviews some recent publications dealing with the phenomenon of polysemy, and addresses some of the questions which they raise. According to a generally accepted definition, polysemy is the association of two or more related senses with a single phonological form. In many respects, the definition is highly problematic. Important foundational questions concern the nature of word senses, how they can be identified, enumerated, and characterized, the manner in which they may be related, and the psychological reality of these constructs. A further question concerns the kinds of linguistic units that are candidates for a polysemy analysis. Also not to be overlooked is that fact that the phonological pole of a linguistic unit is likely to exhibit variation no less than the semantic pole. In spite of the many theoretical and descriptive problems associated with polysemy, it is remarkable that speakers of a language are rarely troubled by it. The paradox is traced back to way in which polysemy is conceptualized by linguists, against the backdrop of ‘idealized cognitive models’ of language. The article concludes with some observations on a usage-based approach to issues raised.

Seeing as though

In this article we address a hitherto unstudied causal conjunction in English, seeing as though . Occurring predominantly in informal registers, the conjunction is typically used to introduce information which the speaker takes to be self-evidently true and on whose basis some further comment, or query, is made. Drawing on data derived from internet searches we draw up a semantic profile of the expression in comparison and contrast with other reason connectives, namely , seeing (that) and since . The data suggest that seeing as though is associated with highly subjective construals of the reason relation. We also address the internal structure of the expression. The use of seeing in a reason conjunction is traced to a common conceptual metaphor, whereby knowing is seeing . More puzzling is the occurrence of as though . While rejecting the possibility of a compositional analysis of the expression, we note that as though is only one of a number of items which can occur with causal seeing . These items have to do with the appearance of things and are in fact able to occur as complementizers after predicates of seeming and appearing. To this extent, as though is consistent with the subjectivity associated with the complex conjunction. In the course of our investigation, we also document the extraordinary proliferation of reason connectives that involve lexical items such as seeing , as , though , and several others, and suggest that this exuberance of new forms may not be unrelated to the subjectivity inherent in the construal of causal relations.

Things and places: The case of the Zulu locatives

Abstract This paper addresses some semantic and syntactic aspects of Zulu locatives. Practically every noun (and pronoun) in Zulu can be locativised. The semantic effect of locativisation is to convert a thing-concept into a place-concept. In many Bantu languages, locatives are fully-fledged nominals; as such, they can function as subjects and direct objects, and control the full range of concordial agreements. Zulu locatives, however, fail to behave like regular nominals. At the same time, the locatives cannot be assimilated to any of the other syntactic categories that are standardly recognised, such as prepositional or adverbial phrases. In view of the fact that the locatives designate places, it is proposed that the locatives should be recognised as a distinct syntactic-semantic category of place-referring expressions.

Prototype Theory in Linguistics

This article is reproduced from the previous edition volume 13, pp. 8954–8957,

Modelling the Emergence of Speech and Language Through Evolving Connectionist Systems

This chapter presents the hypothesis that language knowledge evolves in the human brain through incremental learning and that the process can be modelled with the use of evolving connectionist systems — a recently introduced neural network paradigm. Several assumptions have been hypothesised and proven through simulation: (a) the learning system evolves its own representation of spoken language categories (phonemes) in an unsupervised mode through adjusting its structure to continuously flowing examples of spoken words (a learner does not know in advance which phonemes there are going to be in a language, nor, for any given word, how many phonemes segments it has); (b) learning words and phrases is associated with supervised presentation of meaning; (c) it is possible to build a ‘life-long’ learning system that acquires spoken languages in an effective way, possibly faster than humans, provided there are fast machines to implement the evolving, learning models.